Agitation-Aeration in Submerged Fermentation
II. Effect of Solid Disperse Phase on
M. R. BRIERLEY'
Indutstrial Bioche?nistr-y Section, College of Science and 7'echnology, M1anchester, England
Received for puiblication July 21, 1958
For the successful operation of aerobic fermentations
it is essential to supply the microorganiisms with suffi-
cient oxygen to meet their requirements at any stage
in the process. Failure to supply enough oxygen may
lead to undesirable changes in enzymatic makeup
(Rolinson, 1952) or death of the organisms (Hromatka
et al., 1951), with a subsequent lower yield of the desired
product. In general, the supply of oxygen to bacterial
cultures is more easily maintained than is the supply to
submerged cultures of filamentous fungi. The extremely
rigid (non-Newtonian) broths produced by filamentous
fungi promote the coalescence of bubbles with con-
sequent reduction in transfer area between the gas and
liquid phases and alteration of turbulence and liquid-
film conditions. For example, Chain and Gualandi
(1954) observed about 57 per cent reduction in the
oxygen solution rate in the presence of 1.35 per cent
dry weight of mycelium of Penicillium chrysogenum.
Deindoerfer and Gaden (1955) reported a 100-fold
increase in the rigidity of penicillin broth during the
course of a fermentation in which about 1 per cent dry
weight of mycelium was produced; a mycelial coIn-
centration of 1.34 per cent reduced the oxygen ab-
sorption rate by 85 per cent. Also, the addition of 0.25
per cent antifoam agent (3 per cent Alkaterge C in
lard oil) reduced the oxygen absorption coefficient
by 50 per cent. In studies using paper pulp to simulate
mold mycelium, Bowers (1955) found that the addition
of 2 per cent pulp reduced the oxygen solution rate
given by a sintered aerator by a factor of 20; this was
altered to a factor of 2 by high speed agitation (vaned-
disc impeller) over an open-pipe sparger.
Although the physical effects of mold mycelium on
the oxygen absorption rate in submerged fermentations
have been recognized, a study of the effects of mycelium
concentration and morphology on the correlations
between oxygen absorption coefficients and operating
variables (air flow rate, agitator speed) seemed de-
sirable. Hence studies were made on the effects of
filamentous mycelium, paper pulp, and sago pellets
' Present address: Department of Biological Sciences, Wye
College, near Ashford, Kent, England.
2 Present address: The Upjohn Company, Kalamazoo,
Michigan.
Oxygen Absorption in a Fermentor
AND R. STEEL2
(the latter to simulate smooth mold pellets) on the
rate of solution of oxygen in a fermentor.
M\iATERIALS AND MNETHODS
Fermentor and accessories. Fermentor design (2.5-L
operating capacity) and the polarographic system used
for the determination of oxygen absorption coefficients
(KLa values) have been described previously (Steel
and Brierley, 1959). Briefly, the electrode system
consisted of a rotating "radial" platinum cathode and
a silver/silver chloride reference anode. The applied
potential was -0.90 v. Oxygen absorption-coefficients
were determined by the "gassing-out" procedure of
Bartholomew et al. (1950) and Wise (1951).
Cultivation of a filamentous mold. A strain of Asper-
gillus niger, NIE, was grown in a Waldhof fermentor
(8-L operating volume) in a medium containing sucrose,
salts, and cornsteep liquor. At 5 days, the mycelium was
harvested by vacuum filtration through nylon, washed
3 times by resuspension in distilled water, and finally
washed with sodium azide solution to poison respira-
tory enzymes. The bulk of the water was removed
from the mycelium by drying on a vacuum filter. The
mycelium was used in experiments as soon as possible
although storage at 2 C for 7 days as a wet pad did not
seem to affect its physical properties. Dry weight de-
terminations were made on all samples of mold by
drying to constant weight in a vacuum oven (60 C).
Preparation of paper pulp. The desired amount of
finely shredded filter paper was soaked in the potas-
sium chloride solution (0.5 M) in the fermentor for 30
min prior to experimental w ork.
Preparation of sago pellets. A weighed amount of
dry sago was added to boiling water and stirred until
swelling was complete. The pellets were washed several
times with cold distilled water by decantation and
excess water was removed by vacuum filtration.
Scope of the work. Preliminary studies were made to
determine the optimum positions of sparger and im-
peller within the fermentor since it was believed that
these might be affected by the presence of mycelium.
In subsequent work, oxygen absorption coefficients
were measured in the presence of various amounts of
filamentous mold mycelium, paper pulp, or sago
57
58 5M. R. BRIERLEY AND R. STEEL [VC)L. 7
pellets at agitator speeds from 500 to 1500 rpm and oxygen absorption coefficients within an experimenit,
air flow rates from 0.5 to 5.5 L per min (0.2 to 2.2 under constant operating conditions, was similar to
volumes of air per volume of medium per min). Data that found previously (Steel and Brierley, 1959) for
obtained for a two-phase (gas-liquid) system, over a the gas-liquid system (coefficient of variation less than
similar range of operating variables, have been given 2 per cent). Further, using sago pellets the variability
previously (Steel and Brierley, 1959) and some of these between experiments carried out on different days was
are reproduced in this paper for purposes of com- of a similar order. However, in the case of filamentous
parison.
RESULTS
Operation of the electrode system. In the presenice of
filamentous mold or sago pellets the variability of
TABLE 1
Effect of variouts positions of sparget and impeller on
oxygen absor ption in the absence and presence of
myceliium (0.15 per cent dry weight)*
Oxygen Absorption Coefficient, KLa, hr-
Distance from base of fermentor, in.
Impeller
Sparger
2 24 3½

/7 g 315t 320 320

(174)t (178) (175)
094 310 320 320 1.0 P...,1
2.0
(182) (178) (167) PULP CONCN 4[v
23 --§ - 315
(- ) (167) (175) Figure 2. Effect of paper pulp concentration on the rate of
oxygen absorption. Operating conditions, agitator speed, 800
*
Operatiing conditions; agitator speed, 800 rpm; air flow rpm; air flow, 2.5 L per min.
rate, 2.5 L, per min.
t Mlold absent.
t Mold present.
§ Not measured.

300

250

200

'cyJIS
.
*\

100
0
1 a A -A I

SO

0 0.1 0.2 Q3 GA054 D 3

MOULD CONCN @4
Figture 1. Effect of filamentous mycelium on the rate of
oxygen absorption. Operating conditions, agitator speed, 800
rpm; air flow, 2.5 L per min.
[w/12]
Figure 3. Effect of agitator speed on oxygen absorption
rate in the presence of different amounts of mold. Logarithmic
plot; air flow, 2.5 L per min; the value of Y is given for the
relationship KLa = K N' where N is agitator speed, rpm.
190'9] AGITATION-AERATIO'N IN SUBMERGED FERMENTATION. II 59
mold, the variability between experiments was larger
BOO (up to 10 per cent) than in the absence of mold (about
6 per cent); this larger variation may be attributed to
600 differences in the physical relationship of the hyphae
because of alterations in the degree of physical inter-
lacing of hyphal networks when resuspending the
4001 mycelium in potassium chloride solution in the fer-
mentor (see Deindoerfer and Gaden, 1955). In the
presence of paper pulp the maximum galvanometer
Ic deflection (given by an oxygen saturated solution)
4 200 oscillated over about 4 scale divisions (for example,
-i 78 to 82 mm); also, the scale deflection on aeration
and agitation of a solution previously gassed-out
with nitrogen was not as regular as it was with the
gas-liquid system. Accordingly, more replication (3
100 to 4 runs) was necessary in experiments using paper
80 pulp.
Relative position of sparger and impeller. The positions
60 of sparger and impeller tested are given along with the
results in table 1. The relative position of the two
members made no significant difference in the oxygen
40 solution rate either in the presence or absence of mold.
In subsequent work, the impeller was situated one
400 600 800 1000 2000 impeller diameter (2 in.) from the base of the fermentor
AGITATOR SPEED[N] RPM
and the sparger orifice was located 3; in. below the
Figlure 4. Effect of agitator speed on oxygen absorption rate impeller.
with air flow and paper pulp concentration as parameters. Experiments with mold and with paper pulp. Measure-
Logarithmic plot; the value of Y1 is given for the relationship ment of oxygen absorption coefficients in the presence
KLa = K NY where N is agitator speed, rpm.
of various amounts of mold or paper pulp were made

MOULD CONC t
6001 XZERO

400t

2001 C65
T I
a

100l

801
XsQJI17
,;
60C

I 2 3 4 5 6
I .2 3 4 5 6 7 8 AI R FLOW [G]L/MIN
AIR FLOW [G]L/MIN
Figuire 6. Effect of air flow on oxygen absorption with agita-
Figuire 5. Effect of air flow oxygen absorption rate in the
on tor speed and paper pulp concentration as parameters. Log-
presence of different amounts of mold. Logarithmic plot; arithmic plot; the value of X is given for the relationship
agitator speed, 800 rpm; the value of X is given for the relation- KLa = K Gx where G is air flow, L per min; open markers =
ship KLa = K Gx where G is air flow, L per min. zero pulp; closed markers = 1 per cent pulp.
60 M. R. BRIERLEY AND R. STEEL
under constant operating conditions (agitator speed,
800 rpm; air flow rate, 2.5 L per min). Upon the ad-
dition of mold, the absorption coefficient was reduced
markedly (figure 1). The largest decrease in absorption
coefficient per unit weight of mold was brought about
by the first small additions. Reductions of 67 and 85
per cent in the absorption coefficients were obtained
by the addition of 1 and 2 per cent, respectively, of
paper pulp (figure 2). Low concentrations of pulp
(below 0.5 per cent) did not reduce the oxygen solution
rate as markedly as the addition of a similar weight of
mold. Visual observations of bubble dispersion indi-
cated a smaller bubble area in the presence of the
added solid phase (mold or pulp) because large bubbles
broke at the liquid surface. At the highest concentra-
tion of solid phase tested, movement (mixing) at the
liquid surface was very poor.
The effect of mold concentration on the correlation
between absorption coefficient (KLa) and agitator
speed (N) is shown by the logarithmic plot in figure 3.
The reduction in KLa was less at high agitator speeds
than at low agitator speeds, hence the exponent in
the relation KLa = KNY increased with increase in
mold concentration. Essentially, similar results were
obtained with paper pulp (figure 4) except that the
change in the exponent on N was larger than it was
with the addition of a similar weight of mold.
The correlation between absorption coefficient
(KLa) and air flow rate (G) at constant agitator speed
(800 rpm) decreased with increase in mold concentra-
tion (figure 5). The exponent on G in the relation
KLa = KGX was reduced from 0.53 in the absence of
mold to 0.11 in the presence of 1.73 per cent dry weight
of mycelium. Essentially, similar results were obtained
with paper pulp, the correlation being unaffected by
agitator speeds from 700 to 1500 rpm (figure 6).
Experiments with sago pellets. Under constant con-
ditions of equipment operation (agitator speed, 800
rpm; air flow rate, 2.5 L per min) sago pellets made no
significant difference to the rate of oxygen transfer
since the reduction in KLa in the presence of 3 per
cent sago was only 8.1 per cent (figure 7). No marked
differences in bubble dispersion were detected visually
350_
yJ ~ __
~ .
ax .
LO C 201 30
SAMO CONC N.4~
Figure 7. Effect of sago pellet concentration on the rate of
oxygen absorption. Operating conditions, agitator speed, 800
rpm; air flow, 2.5 L per min.
[VOL. 7
as the sago concentration was increased from 0 to 3
per cent.
DISCUSSION
It is clear from the foregoing results that the mor-
phology of the solid phase largely determines its physical
effect upon the oxygen transfer rate (figures 1, 2, and 7).
Additions of filamentous mold or paper pulp markedly
reduced the oxygen transfer rate because of their effect
on bubble size and liquid turbulence, increasing the
former and decreasing the latter. Although a high
degree of shear, which is desirable for high rates of
oxygen transfer, occurs at the impeller tip, the shear
decreases markedly as the distance from the impeller
tip increases. It was noted visually that relatively little
movement (mixing) occurred at the upper surface of
broth containing about 2 per cent mold. As noted
previously by Bowers (1955), the reduction in oxygen
transfer rate upon the addition of solid phase can be
overcome to a certain extent by increasing the agitator
speed (figure 3).
Smooth pellets which do not tangle together, but
which are easily fluidized, give a system with oxygen
transfer characteristics similar to a gas-liquid system
(figure 7); thus a mold fermentation utilizing smooth
pellets (see Martin and Waters, 1952) seems to offer
a decided advantage as far as oxygen supply is con-
cerned.
What do the results of the present work imply with
reference to the submerged growth of fungi? First,
since the correlation between oxygen absorption rate
and air flow rate decreased markedly as the mold
concentration was increased (figure 5), the practice of
increasing the air flow rate to increase the rate of
oxygen supply to a mold culture becomes almost
valueless. Secondly, if the oxygen solution rate into
broth is 100 mmoles per L per hr, the addition of the
necessary antifoam agent would reduce this to about
50 mmoles per L per hr (from Deindoerfer and Gaden,
1955) and the production of 2 per cent dry weight of
mycelium would further reduce the oxygen supply
rate to about 5 mmoles per L per hr (from figure 1).
For the culture to receive sufficient oxygen its demand
rate must not exceed 5 mmoles per L per hr, an ex-
ceedingly low level. From another viewpoint, in the
present equipment operated at agitator speed 800
rpm and air flow rate 2.5 L per min, the oxygen so-
lution rate was 62 mmoles per L per hr at zero mold
concentration (figure 3). Accordingly, a culture with a
maximum demand rate of 30 mmoles per L per hr
would grow only to a dry weight of 0.3 per cent before
the oxygen supply rate became limiting (from figure
1).
It is apparent from the above examples that further
work is desirable to find better methods of cultivating
molds. It might be possible with other arrangements
1959] AGITATION-AERATION IN SUBMERGED FERMENTATION. II
of baffles and impellers t,o introduce more shear into
the brot,h. Alternatively, studies on submerged mold
growths to obtain cultures showving Newtonian rather
thain non-Newtoni.an char.acteristics are perhaps wvorthly
of consideration.
ACKNOWLEDGMENTS
The authors are grateful to Professor T. K. Walkei
for his interest in this wvork. One of us (M. R. B.)
thanks the Manchester Educat,ion Committee for the
grant of funds, while the other (R. S.) acknowvledges
receipt of an Imiiperial Chemical Industries Research
Fellowship.
SUMMARY
Studies were made on the effects of sparger and
impeller positions, filamentous mold mycelium, paper
pulp, and sago pellets (the latter simulating smooth
mold pellet,s) on the rate of oxygen absorption in a
fermenitor. Wit,hin the range of conditions tested, the
relfative positions of sparger and impeller did not affect
the oxygen solution rate significantly either in the
presence or absence of mold. Marked reduction in
oxygeni solution rate were brought about by the ad-
ditioin of filamentous mold (90 per cent reduction by
2 per cent dry weight of mycelium) anid paper pulp
(85 per cent reduction by 2 per cent dry weight),
whereas the addition of 3 per cent dry weight of sago
pellets made no significant difference. The correlations
between oxygen absorption coefficients and operating
variables (agitator speed, air flow rate) were altered
substanti.ally by the addition of mold or paper pulp to
61
the fermentor; in fact, the practice of increasing the air
flow rate to increase the rate of oxygen absorption was
found to be almost valueless in the presence of mold or
pulp. From the viewvpoinit of oxygen supply rate, the
results of the present work indicate the importance of
mold morphology in submerged fermentations.
REFEIRENCES
BARTHOLoMIEW, W. H., KAROW, E. O., SFAT, M. R., AND
WILHIELM, It. H. 1950 Mlass transfer of oxygen in sub-
merged fermentation of Stiepto;iyces griseus. Ind. Eng.
Chem., 42, 1801-1809.
BOWERS, R. H. 1955 The mechanics of bubble formation.
J. Appl. Chem., 5, 542-548.
CHAIN, E. B. ANI) GITALANDI, G. 1954 Aerationi stuidies.
Rend. ist. stuper. sanita (English ed.), 17, 5-60.
DEINDOERFER, F. H. ANL) GADEN, E. LJ. 1955 Effects of
liquid physical properties on oxygen transfer in penicillin
fermentation. Appl. Microbiol., 3, 253-257.
HRoMATKA, O., EBNER, H., AND CSOKLICH, C. 1951 Un-
terstuchuingen uiber die Essigghnuing. IV. Uber den
'Einfluss einer vollstan(ligen Unterbrechung der B3eliifting
auf die suibmerse Garting. Enzymologia, 15, 134-153.
MIARTIN, S. M. ANI) WATERS, W. R. 1952 Pro(duiction of citric
acid by submerge(d fermentation. ln(l. Eng. Chem., 44,
2229-2233.
ROLINSON, G. N. 1952 Rtespiration of Penirilli?rm chryso-
genuin in penicillin fermentations. J. CGen. Microbiol.,
6, 336-343.
STEEL, R. AND BRIERLEY, M. R. 1959 Agitation-aeration in
submerged fermentation. I. A comparative study of the
sulfite and polarographic methods for measuring oxygen
solution rates in a fermentor. Appl. Microbiol., 7, 51-
56.
WISE, W. S. 1951 The measuirement of the aeration of culture
me(lia. J. (en. Microbiol., 5, 167-177.